Abstract
Biochemical Systems Theory (BST) is a modeling framework that employs power-law formulations to effectively capture the
inherent nonlinearities and heterogeneity of biological systems. Recent research has shown that BST models can be
modelled by reaction networks. However, many key results in Chemical Reaction Network Theory (CRNT) rely on the
condition of weak reversibility - a property often absent in reaction networks derived from BST models. To address this
challenge, this paper develops algorithms for constructing weakly reversible realizations of two variants of BST models:
S-systems and General Mass Action (GMA) systems. By applying these algorithms, fundamental network properties are
simplified, and recent CRNT results regarding the steady states of such systems are validated. Additionally, some of
these algorithms yield deficiency zero networks - a necessary property for the existence of complex-balanced steady
states.
Finally, the proposed algorithms are applied to the GMA representation of the carbon cycle models by Anderies et al. and Heck et al., demonstrating the existence of concentration robustness in these models.
